Meta's latest layoffs begin with a headcount number and end with a capital-expenditure number.
CNBC reported that Meta is slated to start a new round of layoffs this week, with 8,000 projected job cuts, and that more reductions are expected later this year. The company also scrapped plans to fill 6,000 open roles, according to a memo CNBC cited. [1]
The same report says Meta is lifting its 2026 capital-expenditure guidance by as much as $10 billion, reaching as high as $145 billion, as it ramps up artificial-intelligence investment. [1]
That is the story. Not the mood in Menlo Park. Not another essay about Mark Zuckerberg's management style. Labor is being reduced while compute is being funded.
The divergence is efficient. MSM's strongest version is operational: morale, timing, business units, and the difference between the apologetic 2022 layoff note and today's harder tone. X's strongest version is class politics: workers are being removed to pay for machines that may remove more workers later.
The paper's useful middle is the balance sheet. Meta told employees the reductions were part of running the company more efficiently and offsetting other investments, CNBC reported. [1] That sentence should be read like a capital-allocation note. The company is not merely shrinking. It is swapping one kind of capacity for another.
Wall Street does not automatically reward that swap. CNBC's Morning Squawk noted that a CNBC analysis found most companies in a basket of AI-related layoff announcements had seen their stocks drop since the cuts. [2]
That finding should slow the easy version of the story. Investors like efficiency in the abstract, but an AI layoff is not automatically efficient if the new capital base cannot produce durable revenue. A company can remove payroll expense and still increase risk if it commits itself to massive data-center spending before products, ads, assistants, or developer tools prove they can carry the load. [2]
Meta's scale makes the trade especially visible. Eight thousand expected cuts are not a symbolic trimming at the edge of a company. Scrapping 6,000 open roles is not just a hiring pause. Those numbers tell current employees that the old growth machine has been replaced by a narrower test: fewer people, more compute, harder internal standards, and a management theory that AI investment will justify the disruption. [1]
X will understandably turn that into a replacement story. Some of it may be true. But the stronger claim is not that every laid-off worker has been directly replaced by a model. It is that AI has become the capital project against which labor is now measured. Workers do not need a chatbot to take their exact task for AI to change their bargaining position.
The next receipt is not a slogan about efficiency. It is whether revenue, margins, and AI product uptake justify cutting people while buying infrastructure. If they do, Meta will call it discipline. If they do not, the severance line will look less like strategy than a down payment on faith.
-- THEO KAPLAN, San Francisco